One of the most controversial elements of horror in the past few decades has been the use of jumpscares: sudden, loud attacks on the audience. They have been deemed cheap, lazy, and overall just a nuisance. But as a diehard horror fan myself, I’ve realized the error is not with the jumpscare — its with how people use them.

The Anatomy of a Good Jumpscare

Believe it or not, good jumpscares do exist. They’re just hard to pull off.

It’s true — industry plants have found how easy it is to pull the device in order to trick audience into thinking they’ve been scared, thus thinking their work is good horror. But the key in understanding why jumpscares are good is that jumpscares themselves are not what is scary — they are simply a way to damage the audience.

Horror is good because it is the only genre that, across all its mediums, involves direct audience contact. Even if the story has its own protagonists, its the audience that feels the primal fear associated with its events. Jumpscares are a way of manifesting that fear in a real way. It is essentially the same as getting stung by an electrical buzzer — a mild annoyance, but one you certainly wish would never happen again.

Designing good jumpscares is indeed a challenge, but I believe it can be done by following three pillars: priming, punishment, and uncertainty.

Priming

Cry of Fear uses jumpscares in a genius way. Across the entirety of the game, can you guess how many jumpscares the game has?

It has one.

The thing is that this one jumpscare plays at the very beginning of the game. And if you are a well-versed gamer, you’ll immediately strap in knowing you’re going to be playing a jumpscare-filled experience.

And then the jumpscares never come.

What Cry of Fear has done is primed you for the experience. By setting a jumpscare at the very beginning of the game, it has put you in a permanent on-edge state and thus has to try a little bit harder when it comes to its other horror mechanics. This is essentially the smart people way to do what horror hacks do not properly understand (and thus just annoy people with). Jumpscares are always something that should be done as sparingly as humanly possible, in order to keep their appeal. This priming method is pretty much the best way to do that.

Punishment

Horror, particularly in videogames, has a rather unique problem: the fear revolves around not knowing what will happen to you when the monster catches you. This fear immediately dissipates when you realize all that happens is a low-poly face model screaming at you followed by a bloody ‘Game Over’ screen.

Remember earlier, when I said jumpscares were a form of “damaging” the audience? You can take this literally if you make it a game failure state. Five Nights at Freddy’s is a great example of this — if you’re really good at the game, you’ll never be jumpscared. There are no forced scare moments, because the way it utilizes its jumpscares is strategic.

Uncertainty

But horror that is too dependent on one mode will lead to a tolerance in its audience. Fail enough times, and a player will get used to seeing the jumpscares. There just won’t be that hard-hitting impact anymore.

That is, unless you keep things uncertain.

In order for this to work — and thus not be a forced attempt — it has to be used as a positive, rather than a negative. For example, a player might enter a fail state and, wincing, prepare for the jumpscare… only for one to never come. Now, they stay on edge because they can never predict when they’ll be scared. It’s a sort of bastardized Pascal’s wager: stay good at the game and never get jumpscared, or fail and maybe get jumpscared. That will keep the tensions high.

A Disappointing Example

One of the most hyped up (and unfortunately mid) horror titles of the past few years was Don’t Scream. The game was notable for genius and inventive uses of UE 5’s post processing effects, to make it look like its budget was about 500x what it actually was. But what it was most notable for was its unique gimmick: the game would take control of your microphone and, if it heard you scream, you lose. Explains the title.

Now, on its surface, this idea demonstrates perfect use of jumpscare mechanics. In this case its directly tying your failure state to your scare reaction in a way that, in most games before, was only psychological and not built into the game end. All that it needed was a good system to release those scares.

Alas, the final product was dull at best, and hilariously bad at worst.

A lot of people blame the failure of Don’t Scream on the randomness of its jumpscares. But that’s not quite right —- randomness is uncertainty, and that’s a pillar of a good jumpscare formula. So what went wrong?

The way Don’t Scream would scare players is this: it would have you traverse around a realistic looking, found footage processed map, with a lot of great ambient effects. Then, every once in a while, it would try to catch you off guard with a jump scare, thus causing you to scream. If you lasted 10 or so minutes without screaming, you won.

The problem with this system was that it was purely *pseudo-*random, and the amount of jumpscares was tied to the game’s countdown timer. When you first start the game, you had only a very small chance of getting a jumpscare. But by the time the timer was close to zero, you were getting one every five seconds.

This is the opposite of a good jumpscare strategy!

What was happening was that players would pass a certain point, and start getting a jumpscare buffet. It would happen to them so often that they would immediately build a tolerance, and the second half of the game would end up being way easier than the first half.

Another issue was with the strategy of the placement and context of the game’s jumpscares. To put it bluntly, there was no strategy. You would be sitting in a gas station and have a demonic bird clip into the room, screaming and pecking at your face. It might catch you off guard, but more in a “what the fuck just happened” sort of way rather than fear. Jumpscares were almost never tied to locations, player states, or anything — they just showed up, and therefore half the time wouldn’t actually work. The player would be turned the wrong way, or a scare wouldn’t load in properly due to the level geometry. It was a mess.

In the aftermath of Don’t Scream, there was a big call to cull jumpscares out of videogames. But, as I’ve hopefully described to you, this is a bad idea. Jumpscares, just like any other design choice, needs to be used well. If you get it right, people will enjoy it.

[Image omitted]